This PDF file contains the editorial “Editorial: Mt. Graham Revisited” for OE Vol. 29 Issue 09

This PDF file contains the editorial “Book Rvw: Optoelectronic Technology and Lightwave Communications Systems," edited by Chinlon Lin for OE Vol. 29 Issue 09

This PDF file contains the editorial “Book Rvw: Laser Microfabrication: Thin Film Processes and Lithography," edited by Daniel J. Ehrlich and Jeffrey Y. Tsao for OE Vol. 29 Issue 09

Over the past five years, interest in optical pattern recognition appears to have increased. Possible factors include the availability of improved 2-D spatial light modulators (SLMs), new correlation filterdesigns, and improvements in interfacing hardware.

Complex ternary matched filters (CTMFs) can be implemented optically using detour phase and a single ternary spatial light modulator or a Mach-Zehnder arrangement and two ternary signal light modulators. In this paper, we present the design of CTMFs that yield high signal-tonoise ratios (SNRs) and show with the help of simulation examples that their performance (in the sense of SNRs) is close to that of optimal matched filters.

Symmetric nonlinear matched filters (SNMFs) involve the transformation of the signal spectrum and the filter transfer function through pointwise nonlinearities before they are multiplied in the transform domain. The resulting system is analogous to a multistage neural network. The experimental and theoretical results discussed indicate that SNMFs hold considerable potential to achieve a high power of discrimination and resolution and large SNR. The statistical analysis of a particular SNMF in the two-class problem indicates that the performance coefficient of the SNMF is about four times larger than the performance coefficient of the classical matched filter. In terms of resolving closeby signals, there seems to be no limit to the achievable resolution. However, artifacts should be carefully monitored.

We compare the performance of the nonlinear joint transform correlator, the nonlinearly transformed filter based correlator, and the conventional linear correlator. The pattern recognition systems are compared in the areas of correlation peak, peak to sidelobe ratio, and correlation width. We show that for some noisy input scenes, the linear correlator produces an autocorrelation peak that is smaller than the maximum cross-correlation signal, resulting in false signal detection. On the other hand, compared to the linear correlator, both types of nonlinear correlators produce better correlation performance.

An inherent implementational limitation of the recently developed minimum average correlation energy (MACE) filters arises from the circular correlation implicit in the discrete frequency domain approach followed. An alternative methodology that uses space domain computations to overcome these limitations is presented. An improved synthetic discriminant function (SDF) that satisfies the conflicting requirements of reduced noise variance and sharp correlation peaks to facilitate ease of detection is developed by employing a unified framework for design. A quantitative evaluation of the performance characteristics of the new filter is conducted and is shown to compare favorably with the well known minimum variance SDF and the space domain MACE filter, which are special cases of the present design.

Distortion invariant pattern recognition in optical correlators is explored using a linear vector space formalism. Special attention is given to filter design and methods for reducing the clutter problem. The following three constraints are shown to play an important role in the filter's performance: (1) maximize energy projected from the training set onto the filter, (2) use of bipolar filters, and (3) use of higher dimensional filters. An example algorithm for constructing a distortion invariant filter based on these constraints is also given.

A new class of composite filters called the linear phase coefficient composite filters (LPCCFs) is introduced. Unlike previous approaches to composite filter design, this method considers the training set selection and the filter design simultaneously. The LPCCF is obtained by summing the training images weighted by complex weights of unit magnitude and linear phase. This paper also presents a way of combining the outputs of a bank of LPCCFs. Computer simulations are also included to quantify the performance of this approach.

Transform ratio ternary phase-amplitude filters (TR-TPAFs) encoding the modulation states 1, 0, — 1 have been investigated in theory and practice. Simulations have demonstrated that TR-TPAFs can be formulated to provide increased discrimination compared with binary phase-only filters. TPAFs have practical advantages over complex-valued matched filters—efficient electronic filter storage and real-time implementation with available devices, such as magneto-optic spatial light modulators (MOSLMs).These are associated with the use of only three discrete modulation levels. In an experimental study, significant increases in correlation discrimination were demonstrated for several test patterns, showing substantial agreement with computer simulations. The effects of an imperfect MOSLM zero-modulation state, present in our experiment, were modeled and investigated. The performance of the TR-TPAFs for different threshold line angles also was investigated.

A distortion invariant correlation plane filter (DICPF) based on the analysis of variance technique used in experimental statistics is used to analyze peak correlation responses from a time sequenced coherent optical correlator and identify multiple input objects rapidly. A mathematical development of the filter for invariance to an arbitrary number of distortions, such as rotation, scale, aspect, or any other distortion that can be quantized, is presented. This statistical correlation plane filtering technique can be used to process output correlation data from an array of simple or composite filters that are presented either sequentially or simultaneously (space multiplexed at the Fourier plane of a coherent optical correlator). Results are presented from computer simulations and experiments using time sequenced simple binary phase-only reference filters (BPOFs) to achieve rotation and scale invariant recognition of multiple objects. Results from experiments using sequences of composite BPOFs that are invariant over a range of target rotations are also presented to demonstrate the general applicability of this technique. These results indicate that the DICPF can correctly identify multiple objects within the correlator's target group more than 90% of the time, even in the presence of clutter and noise (up to 20% noise), while maintaining a false alarm rate of 6% or less using single BPOFs.

We describe a novel method based on optical cross-correlation and autoconvolution for the selective enhancement of image components with specific symmetry. Such a symmetry filter can operate in real time by using nonlinear optical four-wave mixing. Instead of correlating with a set of preselected references, the input picture is correlated with its own rotated version, which is generated by simple optics. Any patterns that are invariant under the rotation will exhibit strong correlation signals.

The experimental results from an acousto-optic image correlator are presented. The performance ofthis system, which is evaluated in terms of the peak-to-sidelobe ratio, is shown to be improved by subtracting off the signal-dependent bias terms. These bias terms can be optically generated using the same acousto-optic correlator. Structured background features in the input scene can also dramatically affect the correlation result. This paper suggests a novel implementation of the difference-squared error algorithm using a modified acousto-optic architecture that can discriminate better than conventional correlation in scenes with structured background features. Experimental results from real-time systems constructed in the laboratory are presented.

A one lens focal length binary joint transform correlator is described. This correlator uses a magneto-optic spatial light modulator, lens, and standard 8-bit resolution CCD camera. Computer simulations and experimental results of the effects of changes in scale, in-plane rotation, out-of-plane rotation, target/reference separation, and multiple targets are discussed. The performance using actual sensor imagery containing clutter is presented.

In this paper the application of joint transform correlator (JTC) techniques for use in automatic target recognition using actual sensor data is addressed. The problem of interest is the detection and classification of objects in forward-looking infrared (FLIR) images. A JTC architecture using a single magneto-optic spatial light modulator and a chargecoupled device camera is tested. Unique techniques for binarizing the JTC input and the subsequent Fourier transform fringe pattern, commonly called the joint transform power spectrum (JTPS), are presented. Computer simulations and experimental results are provided.

We discuss a number of anomalies that can result in the output of a joint Fourier transform correlator. Some of these are caused by the writing of the joint power spectrum onto a spatial light modulator having a grid structure. Others result from the binary recording of the joint power spectrum. These effects can cause spurious output cross-correlation peaks and can affect the strengths of the desired cross-correlation peaks.

Iterative learning procedures can be employed to generate synthetic reference patterns for joint transform correlators. These patterns can be implemented directly on inexpensive spatial light modulators (SLMs) based on liquid crystal television sets. The algorithm takes into account random electronic noise in the system and compensates for SLM distortions. Experimental results demonstrate the advantages of the new approach with two algorithmic examples.

One of the features in neural computing must be the ability to adapt to a changeable environment and to recognize unknown objects. This paper deals with an adaptive optical neural network using Kohonen's self-organizing feature map algorithm for unsupervised learning. A compact optical neural network of 64 neurons using liquid crystal televisions is used for this study. To test the performance of the self-organizing neural network, experimental demonstrations and computer simulations are provided. Effects due to unsupervised learning parameters are analyzed. We show that the optical neural network is capable of performing both unsupervised learning and pattern recognition operations simultaneously, by setting two matching scores in the learning algorithm. By using a slower learning rate, the construction of the memory matrix becomes more organized topologically. Moreover, the introduction of forbidden regions in the memory space enables the neural network to learn new patterns without erasing the old ones.

This paper presents a technique of using data association target tracking in a motion sequence via an adaptive joint transform correlator. The massive data in the field of view can be reduced to a few correlation peaks. The average velocity of a target during the tracking cycle is then determined from the location of the correlation peak. We have used a data association algorithm for the analysis of these correlation signals, with which multiple targets can be tracked. A phase-mostly LCTV is used in the hybrid joint transform correlation system, and simultaneous tracking of three targets is demonstrated.

A new driver has been developed for the magneto-optic spatial light modulator. This driver allows the user to feed video directly to the spatial light modulator. Several characteristics of the spatial light modulator have been measured and a binary joint transform correlator has been built and tested. The results of these experiments as well as a description of the video driver are presented here.

A hybrid digital/optical joint transform correlator that requires only one spatial light modulator has been designed, constructed, and demonstrated. This was accomplished using a novel video rate multiplexing and demultiplexing circuit that alternates frames of video addressing the SLM. Demonstrations using a modified liquid crystal television and a Faraday effect magneto-optic device show that the performance of the single SLM arrangement is essentially the same as that from the traditional two SLM architecture. The circuit utilizes common electronic components and requires a very small amount of power.

A passive optical tap that uses cladding modes in a single-mode fiber to provide continuous adjustability of tapped power, large tapping ratios, low excess loss, and structural compatibility with integrated circuit technology has been fabricated and demonstrated. This prototype suffers excess loss of less than 0.07 dB and yields tapping ratios as large as 36 dB. System performance analysis of an optical bus incorporating the integrated fiber-optic tap is discussed, and comparison to conventional singlemode fiber-optic taps, such as fused couplers, is made.

We have developed optical measurement and control techniques that allow stable, precise adjustment and tracking of Fabry-Perot plate parallelism and spacing over a distance of many fringes with a precision of 20 nm. This extension of the range of optical position sensing has proved useful in controlling a cryogenic, electromagnetically driven, infrared spectrometer for astronomy.

The linewidth measurement on wafers depends on the small local irregularities in the region being tested. Such defects must be first of all detected and then canceled in order to give an estimation of the average linewidth. Image processing based on the Karhunen-Loève transform of multiple acquisitions in different experimental conditions is proposed. Using statistical information compression, irregularities are easily detected and efficient noise filtering is performed.

A shear plate interferometer has been installed in a mount that allows the direction of shear to be rotated. With this instrument, the defocus and third-order aberrations in a laser beam can be measured very quickly, usually in about a minute including setup.

In this two-dimensional color Schlieren system, the typical square or circular Schlieren filters are replaced by a T-shaped filter. Such a filter lets more light pass through and is more sensitive, thus creating a better Schlieren system.